Honeywell
Scientists have managed to generate real random numbers for the first time with the help of a 56-qubit quantum computer.
In a study published in the journal Nature, scientists from JPMorganChase, Quantinuum, Argonne National Laboratory, Oak Ridge National Laboratory, and the University of Texas at Austin, using a 56-qubit quantum computer, demonstrated certified randomness for the first time. This means that a quantum computer generates random numbers and uses a classical supercomputer to verify that these numbers are truly random and that they have been created anew.
This is a very important step on the way to use of quantum computers in solving practical problems that are currently impossible to implement using classical systems. The certified randomness protocol was first proposed by Scott Aaronson, professor of computer science at the University of Texas at Austin and director of the Center for Quantum Information. Together with former researcher Shi-Han Hung, they developed a theoretical framework and analytical support for the experimental demonstration.
«When I first proposed my certified randomness protocol in 2018, I had no idea how long I would need to wait to see it experimentally demonstrated Developing the original protocol and understanding it is the first step toward using quantum computers to create certified random bits for real-world cryptographic applications», — Scott Aaronson notes.
Last year, developers from Quantinuum and JPMorganChase, along with a team of experts from Google, have announced performing tasks on their quantum computers, that were not possible with supercomputers. This achievement became known as quantum superiority. However, translating this power into solving practical problems remained an open challenge.
The solution was achieved by using random cyclic sampling to generate certified randomness. This is very important for many applications in the field of cryptography and data protection.
Classic computers are not capable of generating true random numbers on their own. Because of this, they are usually connected to a hardware random number generator. However, attackers can access this generator and use it to provide the computer with non-random numbers. This is used in particular to break cryptographic codes. A certified randomness protocol prevents fraudsters from doing this even if they gain access to a quantum computer.
The development team remotely accessed up to a 56-qubit quantum computer Quantinuum System Model H2 with captured ions and generated random certified bits In particular, they have implemented a RCS-based certified randomness extension protocol that outputs more random numbers than it has as input.
The protocol consists of two stages. First, the researchers repeatedly loaded the quantum computer with complex problems, requiring it to solve them quickly, which even the most powerful supercomputer in the world cannot do. The quantum computer could solve these problems by choosing only one of many possible solutions randomly.
The second stage involved mathematical certification of randomness using classical supercomputers. In practice, it was demonstrated that randomness cannot be simulated using classical supercomputers. Using classical certification on several advanced supercomputers with a total performance of 1.1×10 18 operations floating point per second (1.1 ExaFLOPS), the team managed to certify 71,313 bits of entropy.
«This work marks an important milestone in quantum computing, demonstrating the solution of a real-world problem using a quantum computer that goes beyond the capabilities of classical supercomputers This development of certified randomness not only demonstrates advances in quantum hardware, but will also be vital for further research, statistical sampling, numerical modeling, and cryptography», — emphasized Head of Global Applied Technology Research and Engineer at JPMorganChase Marco Pistoia.
In June 2024, Quantinuum upgraded its System Model H2 quantum computer to 56 qubits with trapped ions. The H2 improved the existing state of the art in the field by a factor of 100 due to its high accuracy and overall qubit connectivity, which confirmed the conclusion that the result could not be obtained on any existing classical computer.
The results of the study are published in the journal Nature
Source: SciTechDaily